Thursday, 2017-05-04

AndyWojoVIO is a requirement for NovaLink right?00:23
AndyWojoMy P812L shows no VIO capabilities.00:29
AndyWojoNvm... stupid 812L doesn't have PowerVM enabled.00:47
*** chhavi has joined #openstack-powervm01:11
*** thorst has joined #openstack-powervm01:15
*** chhavi has quit IRC02:13
*** dwayne has joined #openstack-powervm02:22
*** thorst has quit IRC02:41
*** chhavi has joined #openstack-powervm02:46
*** tjakobs_ has joined #openstack-powervm02:49
*** tjakobs_ has quit IRC02:56
*** shyama_ has joined #openstack-powervm03:03
*** dwayne has quit IRC03:11
*** shyama has quit IRC03:11
*** dwayne has joined #openstack-powervm03:18
*** thorst has joined #openstack-powervm03:41
*** thorst has quit IRC03:45
*** k0da has joined #openstack-powervm04:05
*** chhavi has quit IRC04:08
*** thorst has joined #openstack-powervm04:12
*** nbante has joined #openstack-powervm04:22
*** thorst has quit IRC04:30
*** arunman has joined #openstack-powervm05:08
*** nbante has quit IRC05:43
*** thorst has joined #openstack-powervm06:27
*** thorst has quit IRC06:32
*** arunman has quit IRC06:52
*** arunman has joined #openstack-powervm07:02
*** thorst has joined #openstack-powervm07:28
*** thorst has quit IRC07:33
*** k0da has quit IRC07:56
openstackgerritShyama proposed openstack/nova-powervm master: Allow vios to be filtered for rebuild
*** thorst has joined #openstack-powervm08:29
*** k0da has joined #openstack-powervm08:43
*** adreznec has quit IRC08:45
*** adreznec has joined #openstack-powervm08:47
*** thorst has quit IRC08:49
*** k0da has quit IRC10:13
*** arunman has quit IRC10:31
*** arunman has joined #openstack-powervm10:33
*** thorst has joined #openstack-powervm10:46
*** smatzek has joined #openstack-powervm10:48
*** thorst has quit IRC10:51
*** k0da has joined #openstack-powervm11:02
*** edmondsw has quit IRC11:16
*** thorst has joined #openstack-powervm11:33
*** svenkat has joined #openstack-powervm11:36
*** jpasqualetto has joined #openstack-powervm11:57
*** edmondsw has joined #openstack-powervm12:17
*** smatzek has quit IRC12:17
*** jpasqualetto has quit IRC12:49
*** chhavi has joined #openstack-powervm12:52
chhavido u agree with this comment12:53
*** jpasqualetto has joined #openstack-powervm12:54
shyama_thorst: efried if we are mixing fileio and fc types we need to go back and change pypowervm code to accommodate that. let me know your thoughts12:55
thorstchhavi: responded.12:58
thorstshyama_: that makes sense.  But I thought the rebuild map had that logic already?  Because it would 'discover' the volumes and if one of the vioses couldn't discover the volume then it was kicked out12:59
*** thorst is now known as thorst_afk12:59
shyama_thorst_afk: ok I had not done that for fileio. may be i should do that13:00
*** jpasqualetto has quit IRC13:05
chhavithorst: in the code _ISCSI_LOOKUP_COMPLETE is  we are marking true, even if the initiator is discovered or not. don't we need to mark it true only when its discovered.13:17
chhavior here we are considering that we need to run the initiator discovery only once,13:18
*** jpasqualetto has joined #openstack-powervm13:18
chhavithorst: if its not there, then it will never be there. do we mean that13:18
chhavithorst_afk: i think i got it :) since its initiator, it get initialized during open-iscsi installation. which should never be changed13:20
chhavithorst_afk: need workflow for this
*** jpasqualetto has quit IRC13:26
*** jpasqualetto has joined #openstack-powervm13:39
*** esberglu has joined #openstack-powervm13:44
*** dwayne has quit IRC13:47
efriedchhavi thorst_afk Let me ask this: is RebuildSlotMap smart enough to skip a VIOS that doesn't have any slots of the particular type it's trying to populate?13:58
efriedI mean, why do we need to filter at all at this level?  Why not just keep all the VIOSes in there, and let RebuildSlotMap figure it out.13:59
esbergluefried: CI is back up. Just need to update the skip list. About to spawn a node so that I can test out 1.1.3 for you14:08
efriedesberglu your skip list change set is +214:09
efriedAnd thanks.14:09
*** tjakobs_ has joined #openstack-powervm14:11
shyama_efried: the vios members should be given by the volume driver is what we were thinkgin14:12
*** jpasqualetto has quit IRC14:12
efriedYou mean ignore the ftsk feed entirely?14:13
shyama_no filter the feed based on the uuids coming from the driver14:16
shyama_if there are mixed types then we provide a union of the vios and then let the is_volume_on_vios filter further14:22
openstackgerritShyama proposed openstack/nova-powervm master: Allow vios to be filtered for rebuild
openstackgerritTaylor Jakobson proposed openstack/nova-powervm master: Add support for file-backed ephemeral disk
shyama_efried: thorst_afk i added one more patch. let me know what you think14:31
efriedshyama_ Right, so why bother filtering at this level at all?  Just let the is_volume_on_vios do that.14:32
*** kylek3h has joined #openstack-powervm14:33
shyama_efried: yeah that was my initial option :) so we will still keep the vios_uuids property?14:43
openstackgerritRavi Kumar Kota proposed openstack/nova-powervm master: Translation files for supported languages
efriedshyama_ Is it necessary for it to be an instance variable?14:49
*** smatzek has joined #openstack-powervm14:53
*** k0da has quit IRC15:14
*** dwayne has joined #openstack-powervm15:23
shyama_efried: we need it to be same for all volume drivers so I added a property15:42
efriedOkay, I'm getting pretty seriously confused here.  Especially about where the VIOS uuid/wrapper is actually being consumed.15:43
thorst_afkdoesn't this just go back to my original proposal15:44
thorst_afkin the volume driver15:44
esbergluefried: Develop pypowervm passed on my manual run using OOT driver15:45
thorst_afkif the vios_uuid passed in != mgmt partition (because that's the only one we have file system access to), don't do the file system lookup15:45
efriedesberglu Thanks.  Let's move forward with 1.1.3, then?  Anyone have anything else they need in?  thorst_afk adreznec ?15:45
adreznecnothing here15:46
shyama_we have 2 options 1. filter the vios in the volume_on_vios method in the driver. the vios_uuids is not needed in this case. 2. add the vios_uuid property in the volume driver and use it to filter in the rebuild flow and also in deploy flow in future(if needed)15:51
shyama_I think option 1 is good for now. efried thorst_afk15:51
thorst_afkI was never opposed to option 1.16:05
thorst_afkefried: ?16:05
efriedYeah, that seems fine to me.  Means we take this out entirely, yes?
thorst_afkI think so16:08
thorst_afkshyama_: please confirm?16:08
shyama_efried: yes16:08
efriedYeah, so I like that idea.  That chunk of code is offensive to mine eyes.16:09
shyama_ok :)16:09
thorst_afkefried: thoughts on
efriedthorst_afk What about it?16:40
openstackgerritShyama proposed openstack/nova-powervm master: Allow vios to be filtered for rebuild
thorst_afkreview plz?16:40
efriedthorst_afk Done.16:42
shyama_efried: thorst_afk updated the changeset
*** arunman has quit IRC17:03
thorst_afkshyama_: I need help.  Put in a comment, but I'm confused why the entire change isn't just a 2 liner17:07
*** smatzek has quit IRC17:51
*** smatzek has joined #openstack-powervm18:22
*** smatzek has quit IRC18:30
*** k0da has joined #openstack-powervm18:56
esbergluefried: thorst: adreznec: Turns out the services are getting started already with that systemd change. Looks like it might be a problem with pypowervm remote stuff18:57
efriedOh, goodie.  We love dealing with that.18:57
esbergluAt least that's what it appears to be from the above18:58
esbergluefried: You have any time to help debug? Otherwise I can start and let you know if I find anything18:59
efriedesberglu Yeah, fact that it says localhost implies that our session config isn't being called out.19:00
efriedIs that happening *just* in the SEA agent?  Or in all the services that use pypowervm?19:00
esbergluYeah all pypowervm, not just sea19:02
esbergluMan it's gonna be weird getting used to the new logging system19:03
*** jpasqualetto has joined #openstack-powervm19:04
*** chhavi has quit IRC19:08
efriedesberglu Do you need help?19:25
esbergluefried: Got distracted by some CI stuff for a second. About to jump back in19:26
esbergluBut I suspect I will19:26
efriedesberglu First thing to look for is the log messages in local2remote.py19:27
efried'Successfully loaded config ...' or 'Setting Session host to: ...'19:28
efriedIf neither is there, you ain't getting local2remote.19:28
efriedooo, I have an idea what might be the issue.19:28
efriedIf I start a service with systemd, does it load up /etc/environment?19:29
efriedIma guess that's the problem.19:29
efriedthorst_afk Hey, can we get this one through?
*** smatzek has joined #openstack-powervm19:31
thorst_afkefried: looking19:33
efriededmondsw How strongly do you feel about ?19:35
efriedesberglu you too ^19:35
thorst_afkefried: +219:35
edmondswefried I was thinking more about not overwriting something that might be different than about efficiency19:37
edmondswbut maybe it being different would be a bug?19:37
efriededmondsw Most likely, yeah.19:37
edmondswI don't feel too strongly about it19:37
efriedk, I'm gonna push it.19:37
esbergluefried: Yeah looks like I will have to specify the env in the systemd files19:39
esbergluAnd cool19:39
efriedesberglu Makes sense.19:39
efriedthorst_afk Hacking rules in networking-powervm appear to be getting inherited from a mainline project, which is forbidding log translation.19:42
thorst_afko boy19:42
efriedDo we a) break away from mainline hacking rules, or b) remove translation?19:42
thorst_afklets fix it then19:42
thorst_afkremove translation19:42
efriedon it.19:42
openstackgerritMerged openstack/nova-powervm master: Get flavor from instance
esbergluefried: yep that was it19:48
efriedesberglu cool19:48
esbergluSeeing if anything else will fail us now19:49
efriedesberglu needs a rebase20:02
efriedProbably fault of 'get flavor from instance'.20:03
efriedesberglu Ditto
esbergluefried: on it20:12
openstackgerritEric Berglund proposed openstack/nova-powervm master: In Tree Backport: Exception handling #1
openstackgerritEric Berglund proposed openstack/nova-powervm master: In-Tree Backport: TaskFlow
*** thorst_afk has quit IRC20:59
openstackgerritEric Fried proposed openstack/networking-powervm master: Remove log t9n, py34->35, pep8, other tox fixes
openstackgerritEric Fried proposed openstack/networking-powervm master: Re-home neutron-lib constants
*** svenkat has quit IRC21:02
*** jpasqualetto has quit IRC21:03
*** thorst_afk has joined #openstack-powervm21:03
edmondswefried CI passed for
efriededmondsw How can you tell?21:05
edmondswefried oh is it more than seeing Jenkins +1 ?21:06
efriedYou watching jenkins?21:06
efriededmondsw Yeah, our CI jobs run tempest tests with our driver.21:06
efriedIt's, like, esberglu's whole job.21:06
edmondswso how do you tell, if it doesn't update the review?21:07
edmondswI assumed our CI was one of the jenkins gate jobs listed in the review as SUCCESS21:07
*** thorst_afk has quit IRC21:08
edmondswefried isn't that the point of the CI, to post results where folks can see them?21:08
efriededmondsw It does update the review.21:08
efriededmondsw See e.g.
efriedUnder "IBM PowerVM CI"21:09
edmondswefried ahhh... k21:09
efriededmondsw It's considerd a "third-party" CI.21:10
edmondswefried right... I just hadn't thought that through apparently21:10
efriedIt doesn't get to vote for nova proper - but it does gate changes in our *-powervm projects.21:10
esbergluedmodsw: Yeah it just takes way longer than the jenkins gating jobs21:10
adreznecYeah, that's because the jenkins gating jobs for the *-powervm projects don't run devstack, just pep8/flake/etc21:11
esbergluThe systemd stuff is going well. Got the remote pypowervm working. I think the only other major change due to that is going to be log publishing21:13
efriedesberglu Yeah, that's gonna take some getting used to.21:15
*** jpasqualetto has joined #openstack-powervm21:15
efriedesberglu sdague sent a note with some pointers to the devstack functions that pull the relevant logs from journald21:15
esbergluYep. Saw that, haven't tried any of it yet though21:15
*** smatzek has quit IRC21:27
*** dwayne has quit IRC22:02
*** edmondsw has quit IRC22:12
*** edmondsw has joined #openstack-powervm22:15
*** edmondsw has quit IRC22:20
*** jpasqualetto has quit IRC22:21
*** tjakobs_ has quit IRC22:44
*** k0da has quit IRC22:57
*** tjakobs_ has joined #openstack-powervm23:08
*** tjakobs_ has quit IRC23:13
*** edmondsw has joined #openstack-powervm23:25
*** edmondsw has quit IRC23:30
*** svenkat has joined #openstack-powervm23:55

Generated by 2.14.0 by Marius Gedminas - find it at!